Real-time in-tube concrete level tracking during concrete-filled steel tubular arch bridge construction using infrared thermography and computer vision

Abstract

Automated remote monitoring of the concrete pouring process in concrete-filled steel tubular (CFST) arch bridges is a challenging task due to long distances, oblique camera angles, and occlusion, which hinder the accurate and continuous tracking of the process using existing computer vision (CV)-based methods. This paper proposed an integrated CV system for real-time, automated tracking and localization of the in-tube concrete pumping level with infrared thermography. The main contributions include: (1) Proposing a PNP-based orthographic rectification method to accurately correct the scale distortion of oblique infrared images for arch bridge structures. (2) Developing an improved Kalman filter method for stably tracking the concrete pumping level in infrared images with a low signal-to-noise ratio. The results show that the proposed system can achieve mm-level accuracy for the scaled model in indoor experiments, and its effectiveness is evaluated for an actual construction process at a distance of a hundred meters.